In principle, animals would be more fit if parents informed their offspring of the environmental conditions they experienced before the conception of their young. Although inheritance of DNA sequence is largely unaffected by the environment, some epigenetic information can be inherited through the mammalian germ line and may therefore carry environmental information between generations. Epidemiological evidence in humans suggests that poor parental diets can predispose their offspring to cardiovascular disease. We sought to test the hypothesis that epigenetic information transferred between generations can provide offspring with molecular "knowledge" of the environmental conditions experienced by their parents. To search for transgenerational epigenetic inheritance, we therefore carried out a screen for genes in mice that respond to paternal diet. Our preliminary data demonstrate that the expression of genes involved in lipid biosynthesis and cell proliferation increased dramatically, relative to a cohort of control offspring, when their fathers were fed a low-protein diet. Here, we propose to pursue a comprehensive investigation of the mechanisms of transgenerational control of gene expression in mammals. We will further explore the ability of diet to affect offspring's phenotype, and we will determine whether transgenerational information is carried by sperm, seminal fluid, or some other information carrier. Together, these experiments constitute an extensive characterization of a novel pathway linking diet to phenotype across generations. These results will have important implications for human cardiovascular disease, cancer, and more broadly for the epidemiology of complex diseases. PUBLIC HEALTH RELEVANCE: This proposal concerns a comprehensive investigation of the possibility of transgenerational control of gene expression in mammals, using a mouse model to remove confounding effects present in human populations. There is some evidence in the literature that dietary variation can indeed affect disease risk in subsequent generations in humans. If demonstrated conclusively, transgenerational memory of diet would imply that longitudinal studies need to address not only whether parents experienced some environment, but also when this experience occurred relative to conception. This would call for a dramatic rethinking of epidemiological studies in complex diseases such as diabetes, heart disease, alcoholism, and cancer. Thus, our identification and characterization of pathways linking diet in one generation to effects on the next generation is of key importance for understanding and eventually treating a wide range of human diseases.